N.D.G. White

Cereal grain and dockage identification using machine vision

Algorithms were written to extract a total of 230 features (51 morphological, 123 colour, and 56 textural) from the high-resolution images of kernels of five grain types [barley, Canada Western Amber Durum (CWAD) wheat, Canada Western Red Spring (CWRS) wheat, oats, and rye] and five broad categories of dockage constituents [broken wheat kernels, chaff, buckwheat, wheat spikelets (one to three wheat kernels inside husk), and canola (rapeseed with low erucic acid content in the oil and low glucosinolate content in the meal)]. Different feature models, viz. morphological, colour, texture, and a combination of the three, were tested for their classification performances using a neural network classifier. Kernels and dockage particles with well-defined characteristics (e.g. CWRS wheat, buckwheat, and canola) showed near-perfect classification whereas particles with irregular and undefined features (e.g. chaff and wheat spikelets) were classified with accuracies of around 90%. The similarities in shape and size of some of the particles of chaff and wheat spikelets with the kernels of barley and oats affected the classification accuracies of the latter, adversely.

Fungal Detection in Wheat Using Near-Infrared Hyperspectral Imaging

Different species of fungi infect grain in the field and storage facilities. Contamination by fungi in grain is detected and quantified by traditional methods, such as microbial incubation and microscopic detection, which are subjective, labor intensive, and time consuming. An accurate and timely detection technique for fungal growth in grain is needed to prevent grain from spoiling and to reduce quality loss. In this study, the potential of near-infrared hyperspectral imaging to detect fungal infection in wheat was investigated. Wheat kernels infected with storage fungi, namely Penicillium spp., Aspergillus glaucus, and Aspergillus niger, were scanned using a hyperspectral imaging system, and a total of 20 image slices at evenly spaced wavelengths between 1000 to 1600 nm were acquired to form a hypercube. A multivariate image analysis (MIA) technique based on principal component analysis (PCA) was used to reduce the dimensionality of the image hypercubes. Two-class and four-class classification models were developed by applying k-means clustering and discriminant (linear, quadratic, and Mahalanobis) analyses. Two-class discriminant classification models gave maximum classification accuracy of 100%, and on average 97.8% infected kernels were correctly classified by the linear discriminant classifier. The four-class linear discriminant classifier correctly classified more than 95% of the kernels infected with Penicillium and 91.7% healthy kernels. However, the discriminant classifiers misclassified the kernels infected with A. niger and A. glaucus.

Storage and drying of grain in Canada: low cost approaches

Most Canadian grain (>70% of harvests) is stored on the farm. High moisture content of grain at harvest rapidly leads to spoilage and occasionally the production of the mycotoxins sterigmatocystin, ochratoxin A, or citrinin. Near ambient drying systems that consist of an electrical fan at the base of a granary blowing air into a plenum beneath a perforated floor under stored grain is relatively economical. Heat produced by the fan can dry grain by 2% moisture content in 2 months at a cost for electricity of Can.

Mechanical damage to soybean seed during processing

0.87/tonne. Hot air dryers are used on wet grain at harvest to rapidly lower 21% moisture content maize to 15% moisture content at a cost of Can.

Soft X-Ray Inspection of Wheat Kernels Infested by Sitophilus oryzae

7.80/tonne for propane plus aeration costs to cool the grain. Procedures for safe storage of grain are outlined and the capabilities and planned use of a new grain storage research facility at the University of Manitoba are discussed.

Comparison of a Neural Network and a Non-parametric Classifier for Grain Kernel Identification

Abstract The effects of seed cleaning and handling on soybean seed germination and physical integrity were determined with changing seed moisture content (m.c.). In addition, storage behavior of seed and loss of storability caused by damage resulting from free-fall from different heights were determined. Six lots of the variety “MACS-13” at three different m.c.s were passed through a vertical bucket elevator, cleaner with grader, and gravity separator and evaluated for mechanical damage, germination, and vigor index. The storage behavior of the lots, at different stages of processing, was studied by performing an accelerated aging test. The effect of free-fall on quality of the seed was studied by dropping six seed lots from four different heights on to cement and galvanized iron floors. The vertical bucket elevator significantly decreased germination and increased splits and seed coat damage. The seed lots at 12% m.c. (dry basis), suffered less loss in seed quality than the lots at 10% or 11% m.c. The storage quality of seed, as predicted by the accelerated aging test, at 12% m.c. was also better than the lots at 10% or 11% m.c. A free-fall of soybean seed from different heights on to the cement floor resulted in greater loss in quality than when dropped on to the galvanized iron floor.

Image Analysis of Bulk Grain Samples Using Neural Networks

The potential of a soft X–ray method (15 kV and 65 .A) to detect internal seed infestations by the rice weevil (Sitophilus oryzae) in Canada Western Red Spring wheat was determined in this study. The infested kernels were identified by the presence of egg plugs and were scanned with a real–time fluoroscope every 5 to 7 d until the adults emerged from the kernels. A total of 57 features using histogram groups, textural features, and histogram and shape moments were extracted from the X–ray images of the wheat kernels. Parametric and non–parametric classifiers, and a 4–layer back propagation neural network classifier were used to identify uninfested and infested wheat kernels using histogram and textural features independently, and using all 57 features together. There was no significant difference between the classifiers for the identification of uninfested and infested wheat kernels. More than 95% of uninfested kernels and kernels infested by larval stages were correctly identified by all the classifiers. Wheat kernels infested by pupae–adults and insect–damaged kernels were identified with more than 99% accuracy by the classifiers.

Mycotoxin formation in hulless barley during granary storage at 15 and 19% moisture content

Abstract The performances of a four-layer backpropagation neural network and a non-parametric statistical classifier were compared for classification of barley, Canada Western Amber Durum wheat, Canada Western Red Spring wheat, oats, and rye. A total of 230 features (51 morphological, 123 colour, and 56 textural) from the high-resolution images of kernels of the five grain types were extracted and used as input features for classification. Different feature models, viz . morphological, colour, texture, and a combination of the three, were tested for their ability to classify these cereal grains. To make the classification process fast, the number of input features were reduced to 60 and 30. A set of features consisting of an equal number of morphological, colour, and textural features gave the best classification accuracies. The neural network classifier outperformed the non-parametric classifier in almost all the instances of classification.

THREE-DIMENSIONAL, FINITE ELEMENT, HEAT TRANSFER MODEL OF TEMPERATURE DISTRIBUTION IN GRAIN STORAGE BINS

Algorithms were developed to acquire and process color images of bulk grain samples of five grain types, namely barley, oats, rye, wheat, and durum wheat. The images were acquired using a video camera and were digitized using a frame grabber board. The images were stored on a personal computer from where they were accessed by an image processing program which extracted over 150 color and textural features. A neural-network-based classifier was developed to identify the unknown grain types. The color and textural features were presented to a back propagation neural network for training purposes. The trained network was then used to identify the unknown grain types. Results showed a classification accuracy of over 90% for all grain types.

Classification of fungal infected wheat kernels using near-infrared reflectance hyperspectral imaging and support vector machine

Abstract Eleven-kilogram parcels of hulless barley ( Hordeum vulgare L. cv. Condor) at 15 and 19% initial moisture content were kept in simulated storage in a Manitoba farm granary for 20 weeks (June 1996–October 1996) to determine biotic and abiotic changes and mycotoxin production. Temperature, moisture content, CO 2 levels, ergosterol content, seed germination, microfloral infection, and the presence of major mycotoxins were monitored. Ochratoxin A, citrinin and sterigmatocystin reached mean levels of 24, 38 and 411 ppb by 20 weeks in the 19% moisture content barley, but were absent in the 15% moisture content barley; no other mycotoxins were detected. Penicillium species and Aspergillus versicolor (Vuill.) Tiraboschi comprised the predominant microflora. The effect of storage time was apparent at both 15 and 19% moisture content for grain temperature, Alternaria alternata (Fr.) Keissler, Penicillium species and Aspergillus versicolor . At 19% moisture content, storage time also affected moisture content, CO 2 level, ergosterol content, seed germination, and mycotoxin production. At 19% moisture content, elevated ergosterol levels at weeks 4 and 8 appear to offer early warning of the appearance of sterigmatocystin at week 12, and of ochratoxin A and citrinin at week 20.